The proposed work will develop a non-invasive interface (MU Drive(c)) able to extract the incremental motor unit (MU) firings during voluntary muscle contractions, and use this information as control sources to drive upper-limb prostheses for amputees. Motor units provide high levels of incremental control for intact limbs and therefore have the potential of providing a more natural and seamless way of restoring motor function for amputees. The Phase I effort will provide the technical merit and feasibility of the proof-of-principle by demonstrating that increments of MU recruitment and de-recruitment from a single muscle can provide, in real-time, a proportional monotonically increasing and decreasing control signal that substantially improves upon present capabilities of non-invasive prosthetic control. Specifically, because the recruitment and de- recruitment of MUs is hierarchically and proportionally related to the force produced by the muscle, the MU Drive(c) control signal will be smoother and more stable when compared to a standard myoelectric control parameter. Accomplishing this proof-of-principle will require a substantial overhaul of existing technology for extracting MU firings from surface electromyographic (sEMG) signals recorded during voluntary contractions. More than 80,000 lines of code and 1,000 functions will need to be modified and tested with new signal processing components to replace the current post-processing functionality of the algorithm with real-time capability. The algorithms will be tested on sEMG data recorded during opening and closing contractions of the hand from at least 15 subjects with intact limbs. The stability and smoothness of the MU Drive(c) control signal will be compared to a contemporary sEMG amplitude-based measure of control from the same data set. Further experiments will establish the proof-of-principle in 15 subjects with trans-radia amputation, to mitigate the risk of possible difficulties in extracting usable MU data from damaged or surgically displaced muscles. Phase II will advance the single control source capability to richer MU Drive(c) control possibilities of multiple muscle sources to achieve real-time, 3 degrees-of-freedom, precision control of a prosthesis. At its core, the MU Drive(c) system will provide a richer control source than available from the filtered sEMG signal. The benefits of this principle will have far-reaching effects in the fields of prosthetics, robotics and rehabilitaion. Furthermore, on account of its inherent non-invasive design, the proposed technology will provide a lower-cost, risk averse alternative to implanted interfaces under development for prosthetic control. Rehabilitation specialists will be able to provide the necessary adjustments to ensure optimal performance of amputees' prostheses without the need for burdensome surgical interventions. Our economical prosthetic solution will ease financial stress on an already burdened health care system and provide immediate access for a greater number of amputees.